PCGF homologs, CBX proteins, and RYBP define functionally distinct PRC1 family complexes

Abstract

The heterogeneous nature of mammalian PRC1 complexes has hindered our understanding of their biological functions. Here, we present a comprehensive proteomic and genomic analysis that uncovered six major groups of PRC1 complexes, each containing a distinct PCGF subunit, a RING1A/B ubiquitin ligase, and a unique set of associated polypeptides. These PRC1 complexes differ in their genomic localization, and only a small subset colocalize with H3K27me3. Further biochemical dissection revealed that the six PCGF-RING1A/B combinations form multiple complexes through association with RYBP or its homolog YAF2, which prevents the incorporation of other canonical PRC1 subunits, such as CBX, PHC, and SCM. Although both RYBP/YAF2- and CBX/PHC/SCM-containing complexes compact chromatin, only RYBP stimulates the activity of RING1B toward H2AK119ub1, suggesting a central role in PRC1 function. Knockdown of RYBP in embryonic stem cells compromised their ability to form embryoid bodies, likely because of defects in cell proliferation and maintenance of H2AK119ub1 levels.

Figure 1. Proteomic analysis of PRC1 family complexes

A Heat map of PRC1-associated polypeptides. 293T-REx cells expressing the Flag-HA-tagged PRC1 subunit indicated at the top of the table were lysed and interacting polypeptides were recovered by sequential FLAG-HA affinity purification (TAP). The final eluate was subjected to Liquid chromatography-mass spectrometry (LC-MS), and the spectrum count of each protein is color-coded and displayed as a heatmap with more intense color representing more abundant species. Polypeptides are grouped according to which PCGF they associate with (referred to as PRC1.1–1.6, as indicated on the left). B IP from 293T nuclear extracts (NE) using antibodies for RING1B, RYBP, PCGF4, PCGF1, PCGF2, and PCGF6. Bound proteins were resolved on SDS-PAGE and detected by western blotting for the indicated antigens. C Schematic depiction of PRC1 family complexes. See text for details.

Figure 2. Different PCGF-containing PRC1 complexes are targeted to distinct genomic loci

A Read density for RING1B, PCGF1, PCGF2, PCGF4, PCGF5, PCGF6, CBX2, H3K27me3, and H2AK119ub1 ChIP-seq libraries at five loci. A cartoon representing the most enriched PRC1 complex detected at each locus is shown on top. The x axis corresponds to genomic locations, with the scale and genomic position indicated at the top of each panel. The y axis corresponds to the ChIP-seq signal intensity. Spliced ESTs and UCSC genes representation at each locus are shown on the bottom. B Percentage of overlapping gene targets among RING1B and five PCGFs. Each row is normalized according to the total number of target genes identified for each protein. A red box highlights the percentage of gene targets of each protein that overlap with those of RING1B. C GO biological process and mouse phenotype analysis of the targeted genomic regions identified in PCGF1, PCGF2, PCGF4, PCGF5 and PCGF6 ChIP-seq, using GREAT (McLean et al., 2010). The x axis (in logarithmic scale) corresponds to the binomial raw P-values.

Figure 3. High density map and correlation analysis

A High density map of all the RING1B SERs that are also bound by at least one PCGF. Each horizontal line represents a separate TSS and for each protein the ± 5 kb region is shown, with the position of the TSS marked by the vertical green midline. The position of SERs for the indicated protein is represented by horizontal segments. B Clustering of the indicated PRC1 components and H3K27me3 according to pairwise Pearson correlation scores of their ChIP-seq signal in the top 1000 RING1B SERs. C Vertical scatter plot representing H3K27me3 read density (reads per kilo-base per million reads) within the top 50 SERs bound by RING1B and each PCGF. The middle line represents the mean and whiskers mark the range. P-values were calculated using one-way ANOVA with Newman-Keuls post-test comparing all pairs of columns. ***, P<0.001, **, P<0.01 when compared to PCGF1, 5, and 6. D Vertical box plot representing the H2AK119ub1 read density within top RING1B SERs that overlap with RYBP SERs (“high RYBP”) or have no overlap with RYBP SERs (“low/no RYBP”). The middle line represents the mean and whiskers mark the range. P-values were calculated as in (C). ***, P<0.001. E Read density profile for different subunits of PRC1.2, PRC1.4, H3K27me3, and H2AK119ub1 at CCND2 and RUNX1. The x axis corresponds to genomic locations, with the scale and genomic position shown on top. The y axis corresponds to the ChIP-seq signal density. Spliced ESTs and UCSC Genes representation are shown on the bottom. F Top, a cartoon representing the proposed complex composition. Bottom, extent of peak overlap in regions occupied by RING1B, PCGF2, PCGF4, or RYBP and CBX2. The degree of local overlap of regions enriched for the indicated PRC1 subunit with CBX2-enriched regions is plotted on the x axis, and their frequency is plotted on the y axis.

Figure 4. Biochemical isolation of PRC1.4 sub-complexes

A Glycerol gradient analysis of FLAG-purified RING1B complexes. Every other fraction from a 15–35% glycerol gradient was further purified on HA-beads and resolved on SDS-PAGE followed by immunoblotting for the indicated antigens. B IP of FLAG-purified RING1B complexes by antibodies against RYBP, PHC3, and control IgG. The bound proteins were run on SDS-PAGE and detected with the indicated antibodies. C Competitive binding assay to test the interaction of FLAG-RING1B with HA-RYBP and HA-CBX8. HA-CBX8 was incubated with FLAG-RING1B in the absence or presence of HA-RYBP and precipitated with FLAG-beads. The bound fraction was resolved on SDS-PAGE and specific proteins revealed by western blot using HA and RING1B antibodies. D Schematic classification of PRC1.2/1.4 complexes.

Figure 5. Effect of PRC1.4 complex composition on chromatin compaction and H2AK119 monoubiquitination

A Complexes enriched for RYBP or YAF2 (“Fractions RYBP/YAF2”), or for CBX and PHC proteins (“Fractions CBX/PHC”) were isolated as indicated in the scheme on the top, and examined for their compositions by immunoblotting using antibodies as indicated (bottom). B Recombinant proteins and protein complexes isolated as in A were incubated with nucleosomal arrays. The resulting chromatin compaction was examined by differential centrifugation on a sucrose gradient. The DNA contents of the resulting fractions were visualized on an agarose gel, followed by ethidium bromide staining. The fraction of enriched nucleosomes of each samples are underlined in red. C The nucleosome-enriched fractions from B were visualized by electron microscopy (Experimental Procedures). D Coomassie staining of reconstituted complexes or recombinant proteins isolated from baculoviruses infected SF9 insect cells. E Reconstituted complexes were used in an in vitro nucleosomal H2A monoubquitinylation assay (Experimental Procedures). The reactions were resolved on SDS-PAGE, followed by immunoblotting. The top two panels show the monoubiquitinated H2A detected by a specific H2AK119ub1 antibody and an H2A antibody, respectively. Components of reconstituted complexes were also shown by western blotting using antibodies against RING1B, RYBP, CBX2, and CBX8, respectively. Note that multiple bands migrate above RING1B, representing ubiquitinated forms of RING1B. Ponceau staining of histones is shown at the bottom. F Effects of RYBP on E3 ligase activity of reconstituted RING1B- PCGF4 complex toward H2AK119. Varying amounts of RYBP purified from insect cells were used in the in vitro H2A monoubquitinylation assay as in E, and H2AK119ub1 was detected by immunoblotting.

Figure 6. Requirement of RYBP for ES cell differentiation and cell proliferation

A Morphology of embryoid bodies (EBs) formed by TT2 cells with mock or RYBP stable knockdown (Supplemental Experimental Procedures). Note the presence of central cavities in control cells, which was only rarely seen in RYBP knockdown cells. B Quantification of the percentage of EBs showing central cavities in control or RYBP knockdown TT2 cells. Mean percentages and standard deviations were obtained from four independent experiments for each group. P-value was calculated using a type I, one direction T-test. The insert shows the efficiency and specificity of RYBP knockdown. C Profiles of expression of pluripotency (top) and differentiation (bottom) marker genes in TT2 cells or EBs after 6 day differentiation, either with mock or RYBP stable knockdown. Expression levels are normalized over those in TT2 cells with mock knockdown. All mean values of expression levels and standard deviations were calculated from three technical replicates. See Supplemental Experimental Procedures for details. D Growth curve of HEK 293 and TT2 cells with RYBP stable knockdown or mock control. Cells were seeded at 10⁵ at day zero, and cell numbers were counted every day. All mean cell numbers and standard deviations were calculated from triplicate experiments. E The efficiency and specificity of RYBP knockdown in NE were shown by immunoblotting, as well as the effect of RYBP knockdown on H2AK119ub1 using acid-extracted histones from HEK 293 cells in D.